*     ****************************************************************  *;
*	This readme file provides information about how to replicate 	*;
*	the results shown in the PM party paper.    			*;
*									*;
*		   		IIA Tests				*;
*									*;
*	IIAtest_rand.do conducts 100 Hausman tests on Model 6 where we	*;
*	drop 10% of the observations at random each time. The log-file	*;
*	is called IIA_rand.log.  Without the Bonferroni correction, we	*;
*	find IIA violations in 24 of the tests.  With the Bonferroni	*;
*	correction, we find IIA violations in 14 of the tests.  We also	*;
*	demonstrate that if you take the mean of the p-values from the	*;
*	100 tests, then you would not see the IIA violations.		*;
*									*;
*									*;
*				Tables 1 and 2				*;
*									*;
*	stata_mxl.do takes europe.dta and produces all the results 	*;
*	shown in Tables 1 and 2.					*;
*									*;
*				   Figure 1a				*;
*									*;
*	stata_mxl_predictprob.do takes europe.dta and produces Figure	*;
*	1a.								*;
*									*;
*				   Figure 1b				*;
*									*;
*	changelp.do takes europe.dta and produces Figure 1b.		*;
*									*;
*			Table 5	in Online Appendix C			*;
*									*;
*	changemed.do takes europe.dta and produces Table 5a. 		*;
*	changepm.do takes europe.dta and produces Table 5b.		*;
*	changepmc.do takes europe.dta and produces Table 5c.		*;
*									*;
*			Figure 2 in Online Appendix D			*;
*									*;
*	compprob.do takes europe.dta and produces Figure 2 in Appendix	*;
*	D.								*;
*									*;
*			Table 3 in Online Appendix A			*;
*									*;
*	r_code provides R code to run a mixed logit model. This code	*;
*	can be used to evaluate triangular, normal, and uniform random	*;
*	coefficients.							*;
*									*;			
*				Robustness Checks			*;
*									*;
*	stata_mxl_robustness.do does a number of things.  First, it 	*;
*	takes europe_all.dta and examines whether our results change	*;
*	if we keep all selection opportunities except those involving	*;
*	nonpartisan PMs and single-party majorities.  Second, it takes	*;
*	westeurope2_CMP.dta and examines what happens when we use CMP   *;
*	data to identify the median party in Western Europe rather than	*;
*	the country experts from the CPDA project. 			*;
*									*; 		
*     ****************************************************************  *;

 										
		